The Federal Reformulated Gasoline (RFG) Program was mandated by the Clean Air Act Amendments of 1990 (Public Law 101-549) to help mitigate near-ground ozone pollution, a principal component of "smog," in the United States. In the lower atmosphere, ozone is produced by chemical reactions involving nitrogen oxides (NOx), a great variety of volatile organic compounds (VOCs), and carbon monoxide (CO) in the presence of sunlight. All three types of ozone-precursor compounds are. emitted by gasoline-fueled motor vehicles, so the control of motor vehicle emissions has been a major emphasis of the nation's effort over several decades to address the problem of ozone pollution.
The RFG program attempts to lower motor-vehicle emissions through re-engineering gasoline blends. For example, the Clean Air Act mandates a specified minimum oxygen content in RFG blends to help reduce emissions of ozone precursors from gasoline-fueled motor vehicles and to reduce the need for some toxic compounds, such as benzene, in the fuel. By itself, conventional gasoline has no oxygen content. Therefore, oxygen-containing chemical additives, called oxygenates, are blended into the fuel.
Implementation of the RFG program has involved controversy about how to determine which RFG formulations meet the various requirements of the program and which do not. The use of oxygenates in RFG is perhaps the most contentious aspect. Methyl tertiary-butyl ether (MTBE) and ethanol are two of the oxygenates most commonly used to meet the RFG program's oxygen requirement. One aspect of the controversy involves the release of toxic compounds into the environment; for example, a phase-out of MTBE has already been mandated in California
because of concern about environmental risks associated with MTBE leakage into drinking water. The other aspect of the controversy, which is the focus of this report, relates to the ozone pollution problem. MTBE and ethanol can affect the amounts and types of ozone precursor compounds emitted from tailpipes of motor vehicles as well as from the evaporation of unburned fuel. Questions persist over which oxygenate is preferable in terms of air-quality impact. This report addresses the potential impact of oxygenates in RFG on the ozone-forming potential of emissions from motor vehicles.
How should regulatory agencies determine if one RFG blend using a particular oxygenate is preferable to another? In attempting to mitigate ozone pollution, the U.S. Environmental Protection Agency (EPA) currently addresses such questions by estimating the mass of VOC emissions resulting from the use of an individual RFG blend. If the estimated mass of emissions exceeds a specified amount, that fuel blend is disallowed. However, a different method for assessing RFG blends has been proposed. Although certain fuel blends, such as those using ethanol, might result in greater amounts of emissions in terms of mass (because of the volatility of ethanol), it is argued that those emissions have a lower ozone-forming potential compared with emissions from other fuel blends. Therefore, the argument goes, EPA's assessment of RFG blends should be based not only upon mass of emissions, but also upon their reactivity (i.e., ozone-forming potential).
To help assess the scientific underpinning for this question, EPA asked the National Research Council to study it independently. In response, the Research Council formed the Committee on Ozone-Forming Potential of Reformulated Gasoline, which has prepared this report. The committee was charged to assess the utility and scientific rigor of evaluating the ozone-forming potential of the emissions resulting from RFG use (i.e., an approach that takes into account not only the total mass of emissions, but also the potential of the emissions to produce ozone). The committee was not charged or constituted to address the design or implementation of possible new regulations based on the ozone-forming potential of RFG blends. In addition, the committee was not charged or constituted to address relevant, but separate, issues about domestic sources versus foreign sources of fuel, relative energy and cost implications for the production of different RFG blends, relative health and global environmental impacts, or the use of renewable versus non-renewable fuels.
In approaching the task addressed by the committee, it is useful to note the context that has led to the RFG program and, thus, to the need for this study. Efforts to reduce ozone pollution in the United States have clearly had a positive impact on our nation's air quality. After accounting for the effects of meteorological fluctuations, data from EPA's Aerometric Information Retrieval System indicate that peak ozone concentrations in 41 metropolitan areas in the United States decreased by about 10% overall from 1986 to 1997 despite growing fuel usage. Nevertheless, ozone pollution remains a problem; in 1997, about 48 million people lived in areas of the United States that were classified as ozone "nonattainment" areas, and promulgation of the new 8-hr National Ambient Air Quality Standard (NAAQS) of 0.08 parts per million (ppm) for ozone is projected to triple the number of counties in nonattainment and to result in extensive nonattainment in rural areas of the eastern United States. The persistence of ozone pollution has sparked a need for innovative approaches to mitigation, and the RFG program is one such attempt.
An assessment of the ozone-forming potential of emissions from motor vehicles fueled by RFG requires information on the types and amounts of emissions from the vehicles. Gasoline-fueled vehicles emit VOCs, NOx, and CO. VOCs from engine exhaust include many different compounds, some of which are not present in the original fuel but are created in combustion. VOCs can also evaporate from a vehicle's fuel system, and are thus independent of combustion. Each type of VOC can react differently in the atmosphere and thus affect the overall ozone-forming potential of vehicular emissions. NOx and CO emissions are generated during combustion and occur only in the exhaust.
In addition to what and how much is emitted, evaluating the ozone-forming potential of various RFG blends involves assessing how reactive the emitted pollutants might be in the chemical processes that form ozone in the lower atmosphere. If the effect of RFG on air quality is large, then the difference between two blends of RFG might be readily discernible. On the other hand, if RFG has a very small effect on air quality, it is likely to be very difficult to identify which of two RFG blends is preferable in terms of air-quality impacts, let alone to quantify these effects reliably.
With both its charge and the context in mind, the committee undertook a review and analysis of relevant data and literature and also considered written and oral statements from numerous experts from the
academic, private, and public sectors. The major findings of these deliberations and analyses are summarized below.
1. Ozone-Precursor Emissions From Gasoline-Fueled Vehicles
Overall emissions of ozone precursors from gasoline-fueled motor vehicles have substantially decreased in recent decades, largely as a result of government mandates and industry's development and use of new emission controls on motor vehicles.
According to EPA estimates for 1997, emissions of VOCs from on-road gasoline-fueled motor vehicles contributed about 26% to the total inventory of VOC emissions from all sources. Correspondingly, on-road vehicles contributed 22% to the inventory for NOx, and 56% for carbon monoxide (CO). These contributions are projected to continue to shrink in the coming years. If correct, this would imply that the potential impact of using RFG on near-ground ozone concentrations will decrease with time. In fact, air-quality models suggest that implementation of the RFG program reduces peak ozone concentrations by only a few percent. Even if the relative contribution of motor vehicles to the current inventory of ozone precursor emissions from all sources has been underestimated (which, historically, has often been the case), the reduction in peak ozone from the RFG program would still likely be less than 10% at most. Although long-term trends in peak ozone in the United States appear to be downward, it is not certain that any part of these trends can be significantly attributed to the use of RFG.
2. High-Emitting Motor Vehicles
A sizable portion of the ozone-precursor emissions from gasoline-fueled motor vehicles appears to be associated with a relatively small number of high-emitting vehicles in the United States.
Emissions tests, tunnel studies, and remote-sensing of tailpipe exhaust suggest that a disproportionately large fraction of motor-vehicle exhaust emissions arise from a relatively small number of high-emitting vehicles. Many such vehicles have improperly functioning catalyst systems because of catalyst deterioration or improper control of the air-to-fuel ratio. In addition, tests performed during the operation of motor
vehicles indicate that a substantial contribution of emissions occurs during a cold start (i.e., before the catalyst system reaches its operating temperature). The committee did not have sufficient information to assess whether vehicles with malfunctioning evaporative-control systems also are important contributors. The great majority of emissions testing of motor vehicles using RFG has been performed on normally functioning vehicles, and there is substantial uncertainty over how RFG affects emissions from high emitting vehicles. Therefore, it is difficult to quantify total motor-vehicle emissions for an entire motor-vehicle fleet and to assess the efficacy of the use of RFG for the full driving cycle.
3. The Use of Reactivity in Assessing the Ozone-Forming Potential of Emissions
The use of reactivity in assessing the ozone-forming potential of VOC emissions has reached a substantial level of scientific rigor, largely as a result of research sparked by policy making in California over the past several decades.
Ozone chemistry involves many thousands of reactions and a similar number of compounds. Not only does ozone formation respond differently to different VOC compounds and different amounts of NOx, it also responds differently in different locations or pollution episodes. Assessment of reactivity is most appropriate for VOC-limited areas (i.e., areas where ozone concentrations are more sensitive to changes in VOC concentrations than to changes in NOx concentrations). It is likely that reactivity factors could be used in those areas to address nonattainment of the new 8-hr, 0.08-ppm NAAQS for ozone, in a manner similar to that used to address nonattainment of the current 1-hr, 0.12-ppm NAAQS. However, it should be noted that in NOx-limited regions, reactivity—as it is currently used—is of limited value with respect to ozone mitigation. Little research has been undertaken on the derivation and application of NOx reactivity.
4. Relative Reactivity as a Means of Comparing Fuels
The most robust reactivity measures for comparing emissions from different sources are the so-called relative-reactivity factors, but they are often uncertain and of limited utility for comparing similar RFG blends.
These factors are formed by taking the ratio of the reactivity of one compound or emission source to that of another, and thereby canceling out many of the uncertainties associated with the calculation of reactivities. Even so, relative-reactivity factors are typically subject to substantial uncertainty. Various studies suggest that the uncertainty in relative reactivity for emissions, such as those arising from motor vehicles, is generally about 15-30% (at the 95% level of statistical confidence). The major contributors to this uncertainty arise from the substantial variability and difficulty in characterizing how different vehicles respond to changes in fuel composition, the limited amount of test data available, and the limited knowledge of how well a vehicle fleet is characterized by the available data. Because the reactivity of emissions from motor vehicles using various RFG formulations tends to be quite similar and the emissions composition so variable, the reactivity approach is sometimes of limited utility.
5. Reactivity of CO Emissions
CO in exhaust emissions from motor vehicles contributes about 20% to the overall reactivity of motor-vehicle emissions.
The contribution of CO to ozone formation should be recognized in assessments of the effects of RFG. If adding an oxygenate to a gasoline significantly changes the amount of CO emitted by the motor vehicle fleet, this would affect ozone formation. Further, as VOC emissions from mobile sources continue to decrease in the future, CO emissions might become proportionately an even greater contributor to ozone formation. The committee did not conclude that the various RFG oxygenates affected CO emissions to such a degree that they substantially altered reactivity comparisons between RFG blends. However, it is important to note that there are substantial uncertainties in how fuel oxygen impacts CO emissions from high-emitters, as well as in the contribution of high-emitters to overall CO emissions.
6. Overall Air-Quality Benefit of RFG
Emissions tests, tunnel studies, and remote-sensing of tailpipe exhaust indicate that RFG usage can cause a decrease in both the exhaust and evaporative emissions from motor vehicles.
In addition to a minimum oxygen content, the RFG program re-
quires gasoline blinds to have a number of other characteristics that are intended to produce lower emissions. Major contributors to decreased emissions appear to be lowering the Reid Vapor Pressure (RVP)1 of the fuel, which helps depress evaporative emissions of VOC, and lowering the concentration of sulfur in the fuel, which prevents poisoning of a vehicle's catalytic converter by sulfur. Overall, it is estimated that use of RFG can result in approximately a 20% reduction in the mass and total reactivity of VOC emissions from motor vehicles. In addition, such blends can lead to reductions in emissions of CO and some air toxics. Despite such emission reductions, however, the overall effect of the RFG program on ozone air quality is expected to be difficult to discern.
7. Effect of Oxygenates in RFG
The use of commonly available oxygenates in RFG has little impact on improving ozone air quality and has some disadvantages.
Although there is some indication that oxygenates decrease the mass of VOC and CO exhaust emissions, as well as their combined reactivity, the decrease, if any, appears to be quite small. Moreover, some data suggest that oxygenates can lead to higher NOx emissions, which are more important than VOC emissions in determining ambient ozone levels in some areas. Thus, the addition of commonly available oxygenates w RFG is likely to have little air-quality impact in terms of ozone reduction.
The most significant advantage of oxygenates in RFG appears to be a displacement of some toxics (e.g., benzene) from the RFG blend, which results in a decrease in toxic emissions. However, not all air toxics are decreased; for example, emissions of formaldehyde are not decreased and might even be increased by MTBE blends of RFG. Although ethanol blends of RFG might not increase formaldehyde emissions, they lead to increased emissions of acetaldehyde.
8. MTBE Blends Versus Ethanol Blends—Exhaust Emissions
The reactivity of the exhaust emissions from motor vehicles operating on
ethanol-blended RFG appear to be lower—but not significantly lower—than the reactivity of the exhaust emissions from motor vehicles operating on MTBE-blended RFG.
Data from emission tests indicate that there is no statistically significant difference (at the 95% confidence level) between RFGs blended with MTBE or ethanol in the mass of VOC or NOx exhaust emissions from motor vehicles. There is also no statistically significant difference between MTBE and ethanol blends in the reactivity of VOC exhaust emissions. No evidence supports the claim that reactivity-weighted VOC emissions from properly operating motor vehicles using RFG with ethanol would be significantly less than those from motor vehicles using RFG blended with MTBE, even if the ethanol-containing fuel had more oxygen than the MTBE-containing fuel. On the other hand, some data indicate that exhaust emissions of CO from motor vehicles using RFG blended with ethanol are somewhat lower than those of motor vehicles using an MTBE-blended RFG. As a result, a small reduction in the reactivity of the combined VOC and CO exhaust emissions from motor vehicles might result from the use of an ethanol-blended RFG over that of a MTBE-blended RFG.
9. MTBE Blends Versus Ethanol Blends—Evaporative Emissions
Both the mass and reactivity (mass of ozone per mile) of evaporative emissions from motor vehicles using ethanol-blended RFG were significantly higher than from motor vehicles using MTBE-blended RFG.
The higher evaporative emissions of the ethanol-blends were likely due, at least in part, to the fact that such blends had an RVP that is approximately 1 pound per square inch (psi) higher than the equivalent MTBE-blended fuel. Moreover, the increase in total reactivity of evaporative emissions from the ethanol-blended RFG far outweighed the small decrease in the reactivity of the exhaust emissions described in Finding 8. As a result, it appears that a net increase in the overall reactivity of motor-vehicle emissions (exhaust plus evaporative) would result from the use of ethanol-blended RFG (with an elevated RVP) instead of MTBE-blended RFG.
10. Reid Vapor Pressure of Ethanol-Containing Fuel
On the basis of Finding 9 above, it appears likely that the use of an ethanol-containing RFG with an RVP that is 1 psi higher than other RFG blends would be detrimental w air quality in terms of ozone.
This conclusion is consistent with the California Air Resources Board's 1998 evaluation that led to its decision against allowing a 1 psi RVP-waiver for ethanol-containing fuels. However, it should be borne in mind that (1) the committees conclusion is based on tests using properly functioning motor vehicles and, thus, might or might not have underestimated the benefits of using ethanol-blended RFG in high-emitting vehicles; and (2), as discussed earlier, the overall impact on ozone of allowing the use of ethanol-containing fuel would likely be quite small in any case.
11. Use of Reactivity to Evaluate RFGS
The committee sees no compelling scientific reasons at this time to recommend that fuel certification under the RFG program be evaluated on the basis of the reactivity of the emission components.
Analyses of available data on emissions from the use of ethanol-blended RFG and MTBE-blended RFGs showed that mass-emissions differences between the two fuels varied on occasion from the differences found by using reactivity as a basis. However, in no case was the fundamental conclusion concerning the choice of one fuel over another on the basis of relative potential air-quality benefits altered by switching from a mass-emissions metric to a reactivity-weighted metric.
12. Models Used to Characterize Emissions from RFG Blends
The models currently used to inform regulatory decision making—by quantifying emissions from motor vehicles that use RFG blends—are problematic.
The current models are based on regression equations developed from data obtained from a limited set of tests on a small sampling of
motor vehicles. Although the Complex and Predictive models are distinct from models used to estimate the mobile source inventory, their capability of reflecting actual emissions needs to be improved. In some cases, algorithms used to develop the regression equations for the models ignore important parameters that can influence emissions. For example, the Complex Model, developed by EPA, does not account for temperature variations when calculating evaporative emissions. The Predictive Model, developed by the California Air Resources Board, excludes consideration of evaporative emissions. Another potential source of error in both models arises from their treatment of high-emitting vehicles. As noted above, a large portion of motor-vehicle emissions come from high-emitting vehicles. However, the emissions from these vehicles are likely to be quite variable and thus difficult to characterize through sampling a small subset of the total population.
13. Opportunity to Track Effects of Phase II RFG Program
The scheduled implementation of Phase II of the federal RFG program in 2000 offers a unique opportunity to track and document the impact of a new ozone-mitigation program. Plans should be made and implemented for an atmospheric measurements program to assess the impact of Phase II RFG on (1) emissions of ozone precursors from the on-road and non-road motor vehicle fleet, as well as ozone-forming potential of those emissions; and (2) the impact of these changes, if any, on ambient concentrations of ozone and its precursors.